Ian Roberts, Prof. of Linguistics, Univ. of Cambridge. The Wonders of Language (2017 1 ed).

p. 27 Middle

Be sure to get your terminology right: phonetics is the study of the speech sounds themselves; phonology is the study of how languages organise speech sounds into structured systems. To the extent that everyone has the same organs of speech (and the same perceptual apparatus, something we did not really go into in the previous chapter), phonetics is the same everywhere. But different languages may well deploy the same sounds differently, so English phonology and French phonology may be, and in fact to a fair extent are, somewhat different.

p. 82 Top

As we’ve seen, phonetics is about the sounds of language, phonology is about how those sounds pattern in linguistic systems [...]

I know that phonetic transcriptions often contain redundant information. E.g., for English, /r/ is a semivowel ([ɹ]), which we don't need to distinguish from a trilled /r/ ([r], which doesn't exist in English).

But I still don't understand why this redundancy warrants phonemic transcription? Isn't it easier to stick to phonetic transcription and stomach the redundancy? I don't know why, but I prefer phonetic transcriptions, as they spare me the extra step of contemplating which phone fits a phoneme in a particular language.

  • 5
    Phonetic transcriptions can never be as detailed as a spectrogram or a recording; there are always details that distinguish one speaker from another, one emotional context from another, etc. No transcription of any sort can get that. So one has to ask what the purpose of a transcription is. And the answer is that this varies enormously. The original purpose of phonemic transcription was to get a minimal phonetic transcription for native speakers; i.e, to represent all and only the sounds the natives contrasted, and to ignore all the differences that the natives ignored.
    – jlawler
    Sep 3, 2018 at 20:38
  • 3
    That's where alphabets come from for many languages. Language learners is a different matter entirely; if you don't already speak the language, you aren't the audience a phonemic transcription is aimed at. As a language learner, your goal could be to learn to read a phonemic transcription aloud and sound like a native, putting in all the automatic phonetic transitions and variations like a native would. To do that, I agree, close phonetic transcriptions are useful. But would you like a dictionary that uses IPA alphabetic order?
    – jlawler
    Sep 3, 2018 at 20:43
  • @jlawler 1. "The original purpose of phonemic transcription was to get a minimal phonetic transcription for native speakers": but why is this necessary in the first place? The native speakers already know flawlessly what they contrast and ignore. So the redundancy in phonetic transcriptions can't bother or harm them.
    – user5306
    Sep 3, 2018 at 20:46
  • @jlawler 2. "But would you like a dictionary that uses IPA alphabetic order?" I think this is a rhetorical question, yes? The answer's no, right? An IPA alphabetic order wouldn't be "ordered"...would [ä] or [ɒ̈] come first?
    – user5306
    Sep 3, 2018 at 20:47
  • 1
    You'd use an arbitrary ordering system, like all alphabets do. The point is that when all phonetic variations are listed in a dictionary, how do you decide how to look something up? Aspirated T would be different from unaspirated or unreleased or tapped T, for instance. Do you look up appear under [ɨ], [ə], [ʌ], or [ʔ]?
    – jlawler
    Sep 3, 2018 at 20:54

1 Answer 1


You are conflating two issues, namely symbolic representation vs. continuous recording; and the question of redundancy. Speech output, the stream of speech sounds themselves, is continuous. You can perform various phonetic computations on those waveforms without any segmentation. The claim is, though, that human language handles everything symbolically, so [præŋk] is a discrete symbolic representation that can stand for innumerable instances of the word "prank" in English. I presume that you don't need to be told why this transduction from continuous physical waveforms to discrete symbolic representations is necessary. As a reminder, though, we don't memorize the acoustics of a word when we learn the word, so we don't speak in a high pitched voice if we say a word that we learned from a female. Language (as opposed to speech) presupposes symbolic representations, that is, transcriptions.

As far as the phonemics / phonetic distinction is concerned, you'll need stronger evidence that you've found to show that Roberts has a particular view of the nature of the surface representation. The distinction between a phonemic vs. a phonetic transcription is kind of a red herring in contemporary linguistics. The more important question is, what is the ontology of these objects? Generative grammar (Roberts being an exemplar of the field) holds that a "transcription" is a mental representation, and not a heuristic device. There are competing non-mentalist theories of language where a transcription is simply an operationally defined output. Classically, the taxonomic approach assumes the now fictitious front-end of a perfect transcriber (exemplified by the late Peter Ladefoged and Ian Catford), who could listen to speech and produce a symbolic recording of what appears in that token – they convert actual speech to sequences of phones (sequences of symbols). You can then perform a distributional analysis on those symbols and arrive at a phonemic transcription.

The generative account does not in principle mandate an information-compression step of removing predictable information, though historically we were wedded to the idea of eliminating redundancy. We were definitively not wedded to the spurious phonemic / phonetic distinction of taxonomic theory. But we were certainly wedded to the idea that the output of the syntax can be modified by phonological rules. I presume you don't need to be told why we need phonological rules. In short, your question contains a false presupposition, at least from the generative perspective.

  • 2
    I hate generativism and all its works, but I agree with the answer—because it's not just generativism, but its superset structuralism that requires the phoneme as a construct (the distributional analysis you refer to). I'm surprised to hear that the distributional approach is regarded as spurious, although I had already worked out for myself it was profoundly leaky... Sep 3, 2018 at 23:48

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.